Accepted Manuscript

Title: Selection of Escherichia coli heat shock promoterstowards their application as stress probes

Author: Joana L. Rodrigues Marta Sousa Kristala L.J. PratherLeon D. Kluskens Ligia R. Rodrigues

PII: S0168-1656(14)00804-9DOI: http://dx.doi.org/doi:10.1016/j.jbiotec.2014.08.005Reference: BIOTEC 6807

To appear in: Journal of Biotechnology

Received date: 21-3-2014Revised date: 24-7-2014Accepted date: 5-8-2014

Please cite this article as: Rodrigues, J.L., Sousa, M., Prather, K.L.J., Kluskens,L.D., Rodrigues, L.R.,Selection of Escherichia coli heat shock promoterstowards their application as stress probes, Journal of Biotechnology (2014),http://dx.doi.org/10.1016/j.jbiotec.2014.08.005

This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.

Page 1 of 44

Accep

ted

Man

uscr

ipt

1

Selection of Escherichia coli heat shock promoters towards their1

application as stress probes2

3

Joana L. Rodrigues1,3, Marta Sousa1, Kristala L. J. Prather2,3, Leon D. Kluskens1, Ligia 4R. Rodrigues1,3*5

6

7

1Centre of Biological Engineering, University of Minho, 4710 - 057 Braga, Portugal8

2Department of Chemical Engineering, Synthetic Biology Engineering Research Center 9

(SynBERC) Massachusetts Institute of Technology, Cambridge, MA 02139, USA10

3MIT-Portugal Program, Cambridge, MA and Lisbon, Portugal11

121314151617

*Corresponding author: Lígia R. Rodrigues18Address: Centre of Biological Engineering, University of Minho, 4710 -19057 Braga, Portugal20Tel: (+351) 253 604 40121Fax: (+351) 253 604 42922E-mail address: lrmr@deb.uminho.pt23

24

25

26

27

28

29

Short running title: Selection of E. coli heat shock promoters30

31

Page 2 of 44

Accep

ted

Man

uscr

ipt

2

31

Abstract32

The mechanism of heat shock response of Escherichia coli can be explored to program 33

novel biological functions. In this study, the strongest heat shock promoters were 34

identified by microarray experiments conducted at different temperatures (37ºC and 35

45ºC, 5 min). The promoters of the genes ibpA, dnaK and fxsA were selected and 36

validated by RT-qPCR. These promoters were used to construct and characterize stress 37

probes using green fluorescence protein (GFP). Cellular stress levels were evaluated in38

experiments conducted at different shock temperatures during several exposure times. It 39

was concluded that the strength of the promoter is not the only relevant factor in the 40

construction of an efficient stress probe. Furthermore, it was found to be crucial to test 41

and optimize the ribosome binding site (RBS) in order to obtain translational efficiency 42

that balances the transcription levels previously verified by microarrays and RT-qPCR.43

These heat shock promoters can be used to trigger in situ gene expression of newly 44

constructed biosynthetic pathways. 45

46

47

48

49

50

Keywords: E. coli; heat shock promoters; microarrays; RT-qPCR; stress probes51

52

53

54

55

56

Page 3 of 44

Accep

ted

Man

uscr

ipt

3

57

1. Introduction58

The heat shock response (HSR) is an essential and universal mechanism to protect 59

cells (Lindquist, 1986). This mechanism is induced by a variety of stress conditions 60

including physicochemical factors such as heat shock, complex metabolic processes and 61

metabolically harmful substances (Arsène et al., 2000; Zhao et al., 2005) like heavy 62

metals (Lindquist and Craig, 1988). The stress factors increase the level of denatured 63

proteins, resulting in a biological activity change that can be damaging. In Escherichia 64

coli exposed to heat shock conditions, there is an increase of the σ32 transcription factor 65

(Grossman et al., 1984) that allows RNA polymerase to recognize the heat shock gene 66

promoters. Consequently, the expression of heat shock proteins (HSP) increases67

(Lindquist, 1986; Zhao et al., 2005). Most HSP are chaperones and proteases (Arsène et 68

al., 2000). Chaperones aid misfolded and unfolded proteins to fold correctly. Proteases 69

degrade the proteins that chaperones cannot fold correctly, thus ensuring that problems 70

created by stress conditions are minimized (Madigan et al., 2009). The E. coli HSR 71

mechanism is, as in other organisms, very complex. When temperature increases up to 72

42ºC, a cascade of events with several key players is initiated. Despite several scientific 73

efforts (El-Samad et al., 2005; Genevaux et al., 2007; Georgopoulos 2006; Guisbert et 74

al., 2004; Ito and Akiyama, 2005; Janaszak et al., 2007; Janaszak et al., 2009; Yura and 75

Nakanigashi, 1999), some steps of this process have still not been completely 76

elucidated.77

In the past decade numerous studies have been carried out using DNA 78

microarrays to unravel the HSR of E. coli (Nonaka et al., 2006; Rasouly et al., 2009; 79

Richmond et al., 1999; Wade et al., 2006; Zhao et al, 2005). These studies evaluated 80

different control and shock temperatures, exposure times and gene expression induction 81

Page 4 of 44

Accep

ted

Man

uscr

ipt

4

methods (temperature increase, σ32 transcription factor increase, IPTG) leading to the 82

identification of several repressed and overexpressed heat shock related genes. Three of 83

the most highly expressed heat shock genes are ibpB, ibpA and dnaK (Nonaka et al., 84

2006; Rasouly et al., 2009; Richmond et al., 1999; Wade et al., 2006; Zhao et al, 2005). 85

However, gene expression and induction ratios obtained in these studies proved to be 86

difficult to compare, mainly due to the different parameters tested.87

The heat shock promoters can be exploited for designing biological systems 88

where a temperature shift can be used to trigger the expression of a particular gene or 89

set of genes and to build new biological functions and complex artificial systems that do 90

not occur in nature (Khalil and Collins, 2010; Rodrigues and Kluskens, 2011). Induction 91

methods like temperature shift are safe, easy to implement and administer (culture 92

handling and contamination risks are minimized) and less expensive (Valdez-Cruz et 93

al., 2010). For instance, the HSR in E. coli has been studied with a special focus on its 94

potential use as a biosensor (transcriptional biosensing). For that purpose, the HSP 95

promoters (dnaK, grpE, lon, clpB, ibpAB, clpPX, fxsA) have been fused to several 96

reporter genes: lux (Ben-Israel et al., 1998; Dyk et al., 1994; Sagi et al., 2003), gfp 97

(Aertsen et al., 2004; Cha et al., 1999; Nemecek et al., 2008; Sagi et al., 2003) and luc98

(Kraft et al., 2007) that allowed detecting the activation of the HSR by the fast 99

occurrence of light. These biosensors enabled the detection of pollution by metals, 100

solvents and other chemicals. Furthermore, they also helped to identify HSP promoters 101

activated by different types of stress and the overload exerted in E. coli by the 102

production of recombinant proteins. The HSP promoters can further have a significant 103

role in another important field of synthetic biology, namely in engineering bacteria and 104

viruses for therapeutic applications (Shankar and Pillai, 2011), like vaccines (Garmory 105

et al., 2003) and gene delivery vectors (Drabner and Guzman, 2001). Besides sensing 106

Page 5 of 44

Accep

ted

Man

uscr

ipt

5

environmental pollution, it is feasible to program bacteria to sense the tumor 107

environment and selectively release a drug to kill cancer cells (Anderson et al., 2006; 108

Shankar and Pillai, 2011).109

This study aims to identify the strongest heat shock promoters in E. coli to be used 110

in synthetic biology approaches to reprogram organisms and trigger the expression of 111

novel biomolecular components, networks and pathways. The promoter’s strength was112

evaluated using microarray data and was further validated by quantitative RT-PCR (RT-113

qPCR) and by the construction and characterization of stress probes.114

115

2. Materials and methods116

2.1. Media and strains117

E. coli K-12 MG1655 (The Coli Genetic Stock Center - CGSC#: 6300, New Haven, 118

CT, USA) was grown in LB medium (tryptone 1% (w/v), yeast extract 0.5% (w/v), 119

sodium chloride 1% (w/v)) at 37ºC/30 ºC and 200 rpm in Erlenmeyer flasks and in a 2 L120

stirred tank bioreactor (Autoclavable Benchtop Fermenter Type RALF, Bioengineering,121

AG, Wald, Switzerland) Bioengineering, Type R’ALF) (section 2.3). Pre-inocula were 122

prepared from frozen glycerol stocks and incubated overnight (~12 h) in LB at 37 123

ºC/30ºC and 200 rpm. Both Erlenmeyer flasks and bioreactor were inoculated with 1% 124

of inoculum. Biomass was monitored in a spectrophotometer at OD600nm. The cell 125

concentrations at which the heat shock should be applied were determined using 126

biomass calibration curves. Samples were collected over time for RNA extraction or 127

fluorescence measurements.128

2.2. Heat shock probes construction129

Page 6 of 44

Accep

ted

Man

uscr

ipt

6

GFP (mut3b) was amplified by PCR from pSB1A2-GFP (Cormack et al., 1996). The 130

promoters were amplified from E. coli gDNA extracted using the Wizard® Genomic 131

DNA Purification Kit (Promega, Madison, WI, USA). Q5® Hot Start High-Fidelity 132

DNA Polymerase (NEB, Ipswich, MA, USA) was used to amplify GFP and all the 133

promoters used. The heat shock promoter sequences were obtained from the NCBI (The 134

National Center for Biotechnology Information) database. The RBS was calculated 135

using the software RBS Calculator v1.1 (Salis, 2011). GFP and all the promoters were 136

cloned using restriction enzymes and DNA T4 ligase (NEB, Ipswich, MA, USA). GFP 137

was cloned in MCS1 of pETduet (Novagen, Darmstadt, Germany), between the BamHI 138

and AscI site. After cloning GFP, the T7 promoter was replaced by each of the three 139

heat shock promoters (selected after microarrays data treatment) that were cloned 140

between the SphI and BamHI site. The SacI site was added downstream of the SphI site 141

to aid colony screening. E. coli ElectroMAXTM DH10B competent cells (F- mcrA 142

∆(mrr-hsdRMS-mcrBC) φ80lacZ∆M15 ∆lacX74 recA1 endA1 araD139 ∆(ara, 143

leu)7697 galU galK λ- rpsL nupG) (Invitrogen/Life Technologies, Carlsbad, CA, USA)144

(Invitrogen Life Technologies) were transformed with the ligation mixture. Colonies 145

were screened by plasmid digestion and E. coli K12 MG1655 DE3 competent cells146

were transformed with the plasmid containing the gene or promoter of interest. Primers 147

(Sigma-Aldrich, St. Louis, MO, USA) and restriction enzyme sites are listed in Table 1.148

149

2.3. Heat shock experiments150

Erlenmeyer flasks were placed in an incubator at 37ºC and 200 rpm. Bacterial 151

growth was monitored by optical density (OD600 nm). Heat shock was applied at a cell 152

concentration of 5 to 6 × 108 cells per mL, a concentration obtained after approximately153

7.5 h of growth. Control flasks were kept at 37ºC, while the ones corresponding to the 154

Page 7 of 44

Accep

ted

Man

uscr

ipt

7

heat shock assay were immersed in a water bath at 45ºC for 5 min. Samples were taken 155

before heat shock and after 5 min of heat shock. These experiments were conducted six-156

fold (two for microarray analysis and four for RT-qPCR analysis).157

In order to fully characterize the transcriptional profiles of the heat shock genes, 158

two experiments were conducted in a 2 L bioreactor (1 L of LB medium) at 200 rpm, 159

30% of dissolved oxygen, and at different control (30 ºC and 37 ºC) and heat shock160

temperatures (45ºC and 50ºC) during different exposure times and different heating 161

rates (Fig. S1.). A very high heat shock temperature (50ºC) was chosen to observe the 162

differences during and after heat shock under more lethal conditions. 163

The experiments with the heat shock probes containing GFP were also performed 164

in Erlenmeyer flasks and the heat shock was applied in a water bath at 45 and 50ºC for 5 165

or 10 min. In each experiment 5 flasks were permanently held at 30 or 37ºC (control) 166

and 5 flasks were subjected to the corresponding heat shock temperature for 5-10 min. 167

Each culture flask contained one of the five following plasmids: T7promoter_pETduet 168

(empty), T7promoter_pETduet_GFPmut3b, ibpApromoter_pETduet_GFPmut3b, 169

dnaKpromoter_pETduet_GFPmut3b, fxsApromoter_pETduet_GFPmut3b. For the 170

plasmids with a T7 promoter, 0.1 mM IPTG was added at induction time zero. Each 171

experiment was conducted three times at different experimental conditions (heat shock 172

temperature and/or duration).173

174

2.4. RNA extraction and DNase treatment175

All samples collected for microarrays and RT-qPCR analysis were immediately176

stabilized with the RNAprotect Bacteria Reagent (Qiagen, Germantown, MD, USA) and 177

the pellets were maintained at -80ºC.178

Page 8 of 44

Accep

ted

Man

uscr

ipt

8

RNA extraction for microarray analysis was conducted using the RNeasy® 179

Protect Bacteria Mini Kit (Qiagen, Germantown, MD, USA), following the180

manufacturer’s guidelines. The Qiagen RNase-free DNase Set (Qiagen, Germantown, 181

MD, USA) was used to eliminate any DNA contamination from the samples.182

RNA extraction for RT-qPCR analysis was performed by the phenol-chloroform 183

method followed by ethanol precipitation, and then the addition of DEPC 184

(Diethylpyrocarbonate) treated water (Khodursky et al., 2003). To eliminate any DNA 185

contamination from these samples, the Turbo DNA-freeTM Kit (Ambion/Applied 186

BioSystems/Life Technologies, Carlsbad, CA, USA) was used according to the 187

manufacturer’s guidelines.188

RNA concentration and quality were determined using a NanoDrop 189

spectrophotometer (ND-1000, Thermo Scientific, Wilmington, DE, USA). RNA purity 190

levels were estimated by the A260/A280 and A260/A230 ratios. Furthermore, RNA integrity 191

was evaluated by running a 1.5% (w/v) agarose gel electrophoresis loaded with 5 μL of 192

RNA sample.193

194

2.5. Microarrays195

Genetic expression levels of samples grown in Erlenmeyer flasks were analyzed by 196

microarray experiments (two control samples – CA and CB - and two heat shock 197

samples – TA and TB). Total RNA samples were sent to an Affymetrix core facility 198

(Instituto Gulbenkian de Ciência, Oeiras, Portugal) where quality-control analysis was 199

carried out using the Bioanalyzer 2100 (Agilent Technologies, Waldbronn, Germany).200

All microarray steps (cDNA synthesis with appropriate random primers (Invitrogen/Life 201

Technologies, P/N 48190-011, Carlsbad, CA, USA), cDNA fragmentation and labeling 202

through incorporation of biotinylated ribonucleotide analogs (ddUTP-biotin), 203

Page 9 of 44

Accep

ted

Man

uscr

ipt

9

hybridization to the GeneChip® E. coli Genome 2.0 Array (Affymetrix, Santa Clara, 204

CA, USA), staining with streptavidin- phycoerythrin and scanning) were done 205

according to Affymetrix protocols. The GeneChip® includes approximately 10,000 206

probe sets for all 20,366 genes present in four strains of E. coli, including the K12 207

MG1655 used in this study.208

2092.6. RT-qPCR210

The primers used for specific amplification (Invitrogen) are described in Table 2211

and were designed using the Primer 3 program (Rozen and Skaletsky, 2000). The size 212

of all amplicons was between 100 and 105 base pairs. Reverse Transcriptase (RT) 213

reaction was performed in a MyCycler™ Personal Thermal Cycler (BioRad, Hercules, 214

CA, USA) and the iScriptTM Select cDNA Synthesis kit (BioRad, Hercules, CA, USA) 215

was used according to the manufacturer’s instructions, except that the reaction volume 216

used was 10 μL. The final RNA template concentration in the cDNA synthesis reaction217

was 0.05 μg/μL, and was calculated using the RNA concentrations as determined by 218

NanoDrop. The reverse primers used in the RT reaction are also described in Table 2.219

qPCR was used to evaluate the transcription levels of the heat shock genes that 220

were selected from the microarray experiments. qPCR was carried out in a CFX96 Real-221

Time PCR Detection System (BioRad, Hercules, CA, USA) using SsoFastTM 222

EvaGreen® Supermix (BioRad, Hercules, CA, USA) as specified by the manufacturer, 223

except that the reaction volume used was 10 μL. Primer efficiency and correlation 224

values (R2) were determined by the CFX Manager Software (Biorad, Hercules, CA, 225

USA). The cDNA dilution used in the qPCR experiments was 1/80. No-reverse 226

transcriptase controls, in which RNA instead of cDNA was used, were included to 227

evaluate the presence of any genomic DNA. Three and two technical replicates were 228

Page 10 of 44

Accep

ted

Man

uscr

ipt

10

performed for each biological replicate from flask and bioreactor heat shock 229

experiments, respectively. 230

231

2.7. Microarrays Data Treatment232

Analysis of the microarray data was conducted using three different software233

programs (Expression Console®, dChip and SAM - Significance Analysis of 234

Microarrays). The Expression Console® software (Affymetrix, Santa Clara, CA, USA) 235

was used to analyze the raw data from the four arrays (CA, CB, TA and TB) and assess 236

the need for data pre-treatment (background correction, normalization and 237

summarization). For data pre-treatment three different algorithms (RMA - Robust 238

Multichip Average, PLIER - Probe Logarithmic Intensity Error Estimation and MAS5 -239

MicroArray Suite 5) were used. dChip software (Li and Wong, 2003) was used for data 240

pre-treatment and to filter the genes in order to obtain the most expressed ones during 241

heat shock conditions. In this case, the algorithm used for data pre-treatment was MBEI 242

(Model Based Expression Index). For background correction the MM (MisMatch) 243

probes where used to correct PM (Perfect Match) probes (PM/MM difference), and 244

normalization was carried out using the invariant set method. Subsequently, data were 245

filtered according to the name "MG1655," since the GeneChip® also includes three246

other pathogenic E. coli strains and several intergenic regions that are not relevant for 247

the purpose of the current work. Fig. 1A. illustrates the roadmap of the microarrays data 248

treatment.249

The genes were also filtered according to their dispersion coefficient in various 250

samples (σ/ ) and to the percentage of presence of the gene in the four above-mentioned 251

arrays. The ratio of standard deviation and mean of gene expression values across all 252

samples should be greater than a certain threshold (0.5 < σ/ <10). The more variable a 253

Page 11 of 44

Accep

ted

Man

uscr

ipt

11

gene is across samples, the larger this ratio will be. Nevertheless, if a gene is mostly 254

absent across samples, this ratio can be large due to a small mean. For these reasons, the 255

gene should be present in more than 40% of the arrays (i.e. at least in two arrays). 256

Following these procedures, 703 genes were found to verify the filtering criteria.257

The filtered genes were used to compare the different samples and to identify the 258

differentially expressed genes. These genes were selected according to their induction259

ratio and absolute 260

difference between the 261

samples . The 262

t-test for paired samples was performed for each gene identified and a p-value lower 263

than 0.05 was assumed to obtain significant results (Cui and Churchill, 2003). The264

differentially expressed genes were grouped by hierarchical clustering strategies to 265

evaluate the different expression profiles.266

The genes identified by the dChip program as differentially expressed (189267

genes) were subsequently tested in the SAM (Significance Analysis of Microarrays) 268

program (S-test) (http://statweb.stanford.edu/~tibs/SAM/). The "two-class paired" test 269

(two groups, and one-to-one pairing between a member of group A and a corresponding 270

member of group B) was chosen with the following parameters: number of 271

permutations = 5000 (at least 1000 permutations are recommended for accurate 272

estimates) and Δ = 1 (delta parameter is used to adjust the false positive rate, that can be 273

decreased at the cost of missing true positives, or increased at the cost of obtaining more 274

false negatives. The number of significant genes and the number of false significant 275

genes will decrease with increasing delta values) (Dziuda, 2010).276

277

2.8. RT-qPCR Data Treatment278

Page 12 of 44

Accep

ted

Man

uscr

ipt

12

RT-qPCR data were analyzed using the 2-ΔΔCT method described by Livak and 279

Schmittgen (2001). In this method, the amount of mRNA for a target gene is normalized 280

to an endogenous control and relative to a calibrator or arbitrary control condition. The 281

housekeeping rrsA gene, encoding the 16S RNA, was used as an endogenous control to 282

normalize the differences in total RNA quantity (Dong et al., 2008; Kobayashi et al.,283

2006). RT-qPCR values were obtained for each gene for both control and heat shock 284

conditions. Afterwards, data were normalized with the 16S CT value and subsequently285

the expression value of each gene under the control conditions (Livak and Schmittgen, 286

2001):287

Equation 1: ΔΔCT= (CT,heat-shock gene-CT,16S) Heat Shock – (CT,heat-shock gene-CT,16S)288

Control. 289

Finally, the expression level for each heat shock gene (E-ΔCT) and the induction 290

ratio (E-ΔΔCT) were calculated. E represents the real calculated efficiency value for each 291

set of primers, which was very close to 100 % (97.8%-100.8%). 292

293

2.9. Fluorescence measurements and data treatment294

The GFP assay was performed by measuring fluorescence intensity with a 295

fluorescence spectrometer (Infinite 200® PRO, Tecan) at an excitation wavelength of 296

485 nm and an emission wavelength of 535 nm. The measurements were performed in 297

Nunc-Immuno™ MicroWell™ 96 well polystyrene black plates (Nunc/Thermo 298

Scientific, Rochester, NY, USA) with 300 uL of culture sample resuspended in PBS.299

The software used in the analysis was the Tecan i-ControlTM version 1.9.17.0 (Tecan, 300

Morrisville, NC, USA). The specific fluorescence intensity (SFI) is calculated by 301

dividing the raw fluorescence intensity by the optical density at 600 nm and allows 302

comparison of samples with different optical density at time zero. 303

Page 13 of 44

Accep

ted

Man

uscr

ipt

13

304305

3. Results and discussion306

3.1. Differential expression of HSP genes and identification of the strongest HSP307

promoters using microarrays308

From t The microarray results of the heat shock experiments conducted in flasks at 309

different temperatures (held constant at 37ºC relative to incubation at 45ºC for 5 min)310

we were able allowed to identify the strongest heat shock promoters (i.e. the most 311

highly expressed genes). After the heat shock, the percentage of transcripts considered312

present increases slightly (≈ 2%) and the transcripts considered absent decrease313

proportionally. This means that the heat shock led to an increase of the expression of 314

some genes, some of which are possibly not transcribed under normal growth 315

conditions. 316

These heat shock promoters were identified after the data pre-treatment 317

(background correction and normalization, see section 2.7). E. coli K-12 MG1655 genes 318

were selected from all the genes present in the Affymetrix array and were further319

filtered and compared to the criteria mentioned before (Fig. 1A). These criteria allowed 320

to only select genes with low variation between the replicates and that are differentially 321

expressed comparing to the control sample. Based on the different criteria (Fig. 1A.), 322

189 genes were selected and classified as differentially expressed. These differentially 323

expressed genes were then hierarchically clustered. As shown in Fig. 1B., the tree is divided in324

two main branches, displaying underexpressed (upper branch) and overexpressed genes (lower 325

branch). All replicates (CA and CB, TA and TB) demonstrate a great affinity with each other 326

(Fig. 1B.) with a Spearman coefficient near 1. When comparing the two conditions (control and 327

heat shock) the coefficient is near -1, demonstrating that there is a significant difference in gene 328

expression between the samples.329

Page 14 of 44

Accep

ted

Man

uscr

ipt

14

Afterwards, these 189 genes were subjected to the t-test and from the 189 genes330

only 56 were considered as being significantly transcribed. However, the t-test is not 331

satisfactory when used in the analysis of data from microarrays with few replicates (as 332

is the case of this work) since it is difficult to estimate the variance (Cui and Churchill, 333

2003). Thus, an adapted t-test, for example, the S- test, should be used However, since 334

the t-test is not satisfactory in the analysis of data from microarrays with few replicates335

as is the case for this work, the S-test was chosen because it prevents the selection of 336

genes with small changes as significant genes and it uses the values of the replicates to 337

estimate the percentage of genes identified by chance (FDR - False Discovery Rate). 338

Using the S-test, from the 189 genes identified as differentially expressed by dChip339

program, only 128 were considered significant with an FDR equal to 0%. The q-value, 340

which is the lowest FDR at which a gene is described as significantly regulated, was 341

also 0. 342

After comparing the samples, the 189 differentially expressed genes were then 343

hierarchically clustered. As shown in Fig. 1B., the tree is divided in two main branches,344

displaying underexpressed (upper branch) and overexpressed genes (lower branch). All 345

replicates (CA and CB, TA and TB) demonstrate a great affinity with each other (Fig. 346

1B.) with a Spearman coefficient near 1. When comparing the two conditions (control 347

and heat shock) the coefficient is near -1, demonstrating that there is a significant 348

difference in gene expression between the samples.349

Table 3 shows the 34 genes that were found to be the most expressed due to the 350

heat treatment (45ºC, 5 min) and which simultaneously met verified the comparison 351

criteria. In Fig. 1B., these genes can be found in the branch called "TA Expression ≈ TB 352

Expression." The microarray results obtained in the current work are consistent with353

Page 15 of 44

Accep

ted

Man

uscr

ipt

15

others reported in the literature (Nonaka et al., 2006; Rasouly et al., 2009; Richmond et 354

al., 2009; Wade et al., 2006; Zhao et al., 2005).355

356

3.1.2. Selection of the appropriate heat shock promoters to trigger gene expression357

In metabolic engineering it is very important to choose the right promoter(s) since many 358

rate-limiting steps may exist in the pathways and accumulation of intermediate 359

metabolites may occur. Therefore, after constructing a novel pathway(s), the expression 360

levels of all the components should be orchestrated to allow fine-tuning of expression 361

levels and optimization of metabolic fluxes (Khalil and Collins 2010; Li et al., 2012). 362

For this reason, it is essential to construct libraries of engineered tandem promoters with 363

varying strengths (Alper et al., 2005; Ellis et al.,2009; Jensen and Hammer, 1998; Li et 364

al., 2012). Depending on the envisaged purpose for the use of HSP promoters, several 365

options can be considered to make an appropriate choice. For example, if building 366

engineered bacteria to overexpress a given compound is foreseen, the promoter of 367

choice will depend essentially on the product concentration needed and its toxicity to 368

the cell, since a low concentration may not be sufficient for the intended purpose, but 369

too high a concentration can damage the cells. Table 3 illustrates three main situations 370

when considering the genes placed directly downstream of the promoters (*): high 371

induction rate and high expression level during heat shock (ibpA, ycjX, ldhA); high 372

induction rate but low expression during heat shock (mutM, mlc, fxsA) when compared 373

to the expression levels of other genes; and medium induction rate and high expression 374

level (clpB, dnaK, groES, lon).375

3.1.2.1. High induction rate and high expression level 376

If the aim is to achieve a very high concentration of a given compound, the chosen 377

promoter has to allow a rapid induction of gene expression, but also a high expression.378

Page 16 of 44

Accep

ted

Man

uscr

ipt

16

In this study (Table 3) and in several of the microarray-based studies to identify E. coli379

heat shock genes, it was found that ibpA and ibpB were the two most induced genes, 380

with a consistently higher induction rate for ibpB (Caspeta et al., 2009; Richmond et al., 381

1999; Zhao et al., 2005). These genes, along with yidE gene, belong to the same382

transcription unit and the σ32 promoter-dependent is upstream of the ibpA gene.383

Moreover, ibpA expression is higher than ibpB expression, both under normal growth 384

conditions and after heat shock treatment. This suggests that the high induction rate of 385

ibpB is not due to a higher expression compared to the gene ibpA, but to its low 386

expression at 37°C (control conditions). Therefore, at 37ºC the ibpB gene is poorly 387

expressed, possibly due to the short half-life of the σ32 factor, while under heat shock 388

conditions, due to an increase in denatured proteins, the half-life of σ32 increases and the 389

ibpB gene is transcribed similarly to the ibpA gene. Hence, despite the high induction 390

rate of the ibpB gene, one can conclude that the promoter upstream of the ibpA gene is391

by far the one that allows a more pronounced expression of the HSP genes (Caspeta et 392

al., 2009). As can be seen in Table 3, the strongest promoter following the ibpA gene 393

promoter is the promoter upstream the operon comprising ycjX, ycjF and tyrR. These394

three genes have been reported to be σ32 dependent in all microarray studies conducted 395

so far. The ldhA gene, encoding the enzyme D-lactate dehydrogenase, was also 396

identified by several authors as a heat shock gene (Richmond et al., 1999; Wade et al., 397

2006; Zhao et al., 2005). The difference in hslJ gene expression, present in the same398

transcription unit, is not so evident in the current results but remains significant.399

3.1.2.2. High induction rate but moderate expression during heat shock400

If a fast expression induction and, at the same time, a more moderate expression is 401

required after heat shock, then mutM, mlc, fxsA and prlC promoters can be used. These402

genes have been reported in several studies as heat shock genes, their functions are 403

Page 17 of 44

Accep

ted

Man

uscr

ipt

17

well-known and their promoters have been properly identified (Nonaka et al., 2006; 404

Shin et al., 2001; Tsui et al., 1996; Zhao et al., 2005). However, in the current work the 405

expression values obtained for mutM and mlc gene duplicates are very different and not 406

conclusive, which can be clearly seen by the t- and S-test results (Table 3). The same 407

occurs with the yibA gene which was only been identified as heat shock gene by Zhao et 408

al. (2005) and Wade et al. (2006), although the sequence of its promoter has not been 409

defined. Thus, if a rapid induction and a lower gene expression are required, then the 410

fxsA and prlC promoters would be a preferable choice. 411

3.1.2.3. Medium induction rate and high expression level412

The promoters associated with other genes with a relatively high induction ratio 413

and expression, such as htpG, ybbN, htpX, yceP, clpB, yrfH, hslV, dnaK, groES, lon, 414

yhdN and sdaA, can also be considered as strong promoters and therefore may be used415

in the construction of engineered bacteria. However, it is necessary to take into account 416

that the lower induction rate in these cases is primarily due to a high expression of 417

genes in control samples. Thus, it is necessary to evaluate any implications that the418

possible expression of these genes could have on the applications envisaged before the 419

heat shock treatment occurs. If the production/expression of the compound/gene before 420

heat shock induction leads to misleading/undesirable results, then these promoters421

should not be used. As can be observed in Table 3, dnaK has the highest expression at 422

45 ºC, followed very closely by ibpA. Although they have similar expression levels 423

under heat shock conditions, the dnaK induction rate is much smaller than that of ibpA. 424

DnaK is one of the most reported chaperones and belongs to the DnaK system, together 425

with the DnaJ (known to assist the DnaK chaperone) and GrpE proteins (nucleotide 426

exchanger factor). In the absence of stress, these proteins have an important role in the 427

regulation of the stability of the σ32 transcription factor, and in the maintenance and 428

Page 18 of 44

Accep

ted

Man

uscr

ipt

18

restoration of the homeostatic balance. As for the ybbN, htpX, yceP, yrfH, hslV, yhdN429

and sdaA genes, it is known that all of them are also associated with σ32 controlled 430

promoters (Chuang et al., 1993; Cowing et al., 1985; Kornitzer et al., 1991; Nonaka et 431

al., 2006; Zhao et al., 2005). The two htpG genes and clpB, groES and lon genes have 432

σ32-dependent promoters but they are also controlled by σ54 and σ70 transcription factors. 433

Therefore, the possibility that the high ratio and expression levels in this study are due434

to the higher affinity of the other factors to the promoter cannot be ruled out.435

Nevertheless, since the binding sites of the two factors are superimposed, the use of 436

these promoters with possibly more affinity for other transcription factors is not 437

necessarily a disadvantage, since both factors can hypothetically bind during heat shock 438

induction, which in turn could lead to a high gene expression.439

3.2. Validation by RT-qPCR of the differential expression of HSP genes440

Validation by quantitative RT-PCR (RT-qPCR) of the overexpression of some of these 441

HSP genes identified by microarrays, for example ibpA, fxsA and dnaK, is required 442

before using their promoters to trigger a gene or a biosynthetic pathway by heat. For 443

RT-qPCR validation, flask experiments were conducted twice in duplicate using the 444

same temperatures and exposure time as for the microarray studies (Fig. 2.). 445

Furthermore, several experiments were conducted in a bioreactor at different heat shock 446

temperatures, during different exposure times and at different heating rates in order to 447

fully characterize the transcriptional profiles of the selected heat shock genes (Fig. 3.).448

Fig. 2A. shows the relative expression of the three selected genes (ibpA, dnaK, fxsA)449

with respect to the rrsA gene in the two sample types (control and heat shock). A 450

different expression of the heat shock genes could be observed when comparing the 451

control with the heat shock sample. In the case of the heat shock genes it was found that452

Page 19 of 44

Accep

ted

Man

uscr

ipt

19

the gene expression is lower in the control, which was foreseen since the heat shock 453

transcripts are lower at 37ºC. Comparison of the three heat shock genes shows that the454

dnaK gene possessed the highest expression in heat shock conditions. However, this 455

gene also showed the highest expression in normal growth conditions, which in turn 456

contributes to a lower induction ratio (Fig. 2B.). The ibpA gene showed the highest 457

induction ratio, since it has a high expression in heat shock conditions but a low 458

expression at 37ºC. The fxsA gene showed a moderate expression under heat shock 459

conditions when compared to the other two heat shock genes, as expected. 460

As can be observed in Fig. 3A. and Fig. 3B., the expression profiles of the three 461

heat shock genes obtained in the bioreactor are very similar. The expression increases462

when temperature up-shift starts and maximum expression occurs when the culture 463

medium reaches 45ºC (Point D, Fig. 3A.) and 50ºC (Point O, Fig. 3B.). After reaching 464

the maximum temperature, the gene expression starts to decrease until a new stationary 465

state is reached due to bacterial adaptation to heat. Also, following the temperature 466

down-shift to the original temperatures, 37ºC (Point I, Fig. 3A.) and 30ºC (Point T, Fig.467

3B.), the heat shock gene expression almost returns to its original basal level. During 468

adaptation, to establish the homeostatic balance, σ32 transcription factor is degraded by 469

proteases, such as FtsH. This occurs due to the chaperones negative modulation 470

mechanism, in which they regulate their own synthesis by controlling the activity, 471

stability and synthesis of the σ32 transcription factor. The HSR at 45ºC was found to be472

in accordance with previous studies conducted at 42ºC (Nagai et al., 1991; Rasouly et 473

al., 2009; Tilly et al., 1989). In contrast, the heat shock response at 50ºC was more 474

prolonged, which was expected since it is more lethal to the cells and it is more difficult 475

for the cells to adapt.476

Page 20 of 44

Accep

ted

Man

uscr

ipt

20

From the analysis of the expression profiles represented in Fig. 3A. and Fig. 3B., 477

we can conclude that the ipbA gene had a higher fold change, since in normal growth 478

conditions (37 ºC) it is less expressed than dnaK gene, and at heat shock temperatures it 479

reaches levels similar to the ones observed for the dnaK gene. As expected, the fxsA480

gene showed a moderate expression. These results are in accordance with the ones 481

obtained in the flasks experiments. Fig. 3C. and Fig. 3D. show the genes fold change for 482

the two experiments after normalization of the data. The gene expression is higher in the 483

experiment conducted at the lethal temperature (50ºC) as for this temperature, the heat 484

shock genes’ transcription is controlled by the transcription level of the rpoH gene. The 485

rpoH gene encodes the σ32 transcription factor. The rpoH gene transcription is486

controlled by different transcription factors depending on the temperature. At 30ºC or 487

37ºC, the rpoH gene transcription is controlled by three σ70 transcription factor 488

dependent promoters (Erickson et al., 1987). In heat shock conditions around 42-45ºC,489

the transcription of the rpoH gene does not increase significantly. In these cases, the 490

increase of σ32, and consequently the expression of the heat shock genes, is due to the 491

de-repression of rpoH gene translation. However, in lethal heat shock conditions 492

(greater than 45ºC) the HSR is also controlled at the rpoH transcription level. When the 493

cells are maintained at these elevated temperatures a balanced growth is no longer 494

possible, the HSP are essentially the only proteins to be synthesized (Raina et al., 1995)495

and the HSR remains elevated until the cells begin to die, unless the temperature 496

decreases (Erickson et al., 1987; Guisbert et al., 2004; Yamamori et al., 1978; Nagai et 497

al., 1991).498

3.3. Construction of heat shock inducible stress probes using GFP499

Since we aimed to identify the strongest heat shock promoters that can trigger a 500

biosynthetic pathway as a result of a temperature increase, besides the transcription 501

Page 21 of 44

Accep

ted

Man

uscr

ipt

21

levels study, it is also important to study and optimize the protein translation to obtain 502

high protein expression. To study the protein translation, the HSP promoter elements 503

were fused to the gfp reporter gene. This strategy has been applied in the past by Cha et 504

al. (1999), where the heat shock response of E. coli resting cells was quantified by 505

dnaK, clpB, and rpoH GFP promoter probes. In their experiment they used chemical 506

(addition of IPTG, acetic acid, ethanol, phenol, antifoam or salt) and physical stresses 507

(heat shock or nutrient limitation) to induce GFP expression. Gill et al. (1998) and Seo 508

et al. (2003) also fused GFP to rpoH, clpB and dnaK promoters to monitor the toxic 509

effect of dithiothreitol addition and compare the cellular stress levels of four strains510

caused by temperature up-shift, respectively. Lu et al. (2005a,b) and Messaoudi et al. 511

(2013) constructed oxidative stress probes by fusing the promoters of several oxidative 512

stress gene promoters with GFP. Other stress probe strategies have also been used. Van 513

Dyk et al., (1994), Bianchi and Baneyx (1999a,b), Lesley et al. (2002), Pérez et al. 514

(2007) Kraft et al. (2007) and Kotova et al (2010) fused heat shock promoters to β-515

galactosidase (lacZ) and to bioluminescent luciferase bioreporters (lux and luc) to 516

investigate the toxicity of heterologous protein expression and other compounds. In our 517

study we used the gfp reporter gene because of its small size, minimal toxicity, high 518

stability and ease of use (March et al., 2003).519

The use of stress probes with the HSP promoters fused to gfp helped to study how 520

efficient the promoter sequences and the original heat shock RBS are in the translation 521

of a gene. Different experiments were done with different controls (30 ºC and 37 ºC) 522

and heat shock (45 ºC and 50 ºC) temperatures and different heat shock duration (5 and 523

10 min). Fig. 4. shows the fluorescence results obtained for the above mentioned 524

experiments. As observed before, the dnaK promoter allows the highest expression after 525

heat shock treatment, but the induction ratio is not as high as for ibpA since the dnaK526

Page 22 of 44

Accep

ted

Man

uscr

ipt

22

expression at 37ºC is already very high and in most cases even higher than ibpA527

expression after heat shock conditions. The fxsA promoter showed less GFP expression, 528

corroborating the microarrays and RT-qPCR results. Comparing the two heat shock 529

temperatures tested (45 and 50ºC) it can be concluded that the highest temperature 530

allows a higher promoter response and consequently, a higher GFP expression.531

However, the fluorescence level at 50ºC depends on the initial temperature (30 and 532

37ºC). In general, an initial lower temperature (30ºC) that corresponds to less basal 533

fluorescence also corresponds to lower specific fluorescence intensity at 50ºC. The SFI 534

obtained with the original T7 promoter (data not shown) was around 33000 and the 535

fluorescence reached the steady state in about 2–3 h (37-30ºC experiments) and 536

remained stable until the end of the experiment. As expected, the T7 promoter is a better 537

option as the expression obtained is higher and more stable. However, the use of IPTG 538

has some disadvantages, such as cost and toxicity, and limited utility for in situ gene 539

expression. The lower stability of the mRNA when using heat shock promoters can be 540

overcome by adding sequences to form stem-loop at its 5’ and 3’ untranslated regions 541

(UTRs) (Chen et al., 1991; Higgins et al, 1993). This prolongs the half-life of a 542

normally unstable mRNA by presumably interfering with RNase binding.543

These results (Fig. 4.) were obtained after studying four different promoter 544

sequences (supplementary material, Table S1 and S2). The first heat shock promoter 545

sequences used to replace the T7 promoter were obtained from gDNA amplification of 546

the -50 region until the start codon (ATG) of the heat shock gene. The TATA box 547

regions, transcription initiation site, RBS and translation initiation site of the native 548

proteins were used for GFP regulation. The only promoter that showed a significant 549

increase in GFP fluorescence after heat shock was dnaK. In order to improve the 550

expression of GFP, the promoter sequence from the -150 region to the heat shock gene 551

Page 23 of 44

Accep

ted

Man

uscr

ipt

23

ATG was amplified, towards the inclusion of some possible existing binding sites for 552

transcription activator proteins present in the region upstream of the core promoter, 553

referred as the upstream (UP) element (Ross et al., 1993). The new sequences did not554

result in an effective improvement of the GFP expression level. Therefore, the promoter 555

sequence from the -150 region to the +21 nucleotides from the three heat shock genes556

was amplified. Several authors (Bianchi and Baneyx, 1999a, b; Pérez et al., 2007) have 557

reported good results when including the first 21 nucleotides of the ibpA open reading 558

frame in a lacZ gene fusion. However, fluorescence did not improve and therefore the 559

experiments were repeated and samples were analyzed by RT-qPCR. The preliminary 560

results showed a significant increase of gfp transcription after heat shock which allowed 561

us to conclude that the lack of fluorescence after heat shock induction was not related to 562

the promoter strength but to the translational efficiency. To overcome this issue, the 563

RBS Calculator v1.1 (Salis, 2011; Salis et al., 2009) software was used. Twenty N 564

nucleotides (any base calculated by the software) were added to the RBS constraints, 565

followed by GGATCC, which corresponds to the BamHI site. The target translation 566

initiation rate was set to 20000 for the three cases. With these new constructions, we 567

were able to observe an increase of fluorescence in the fxsA and ibpA stress probes. 568

However, the fluorescence for the fxsA probe was higher than for ibpA, which was not 569

expected according to the previously observed mRNA levels. The translation initiation 570

rates determined using the RBS calculator for the heat shock gene promoters in the E. 571

coli genome presented the following relation: . To 572

obtain GFP expression results more similar to the ones of the heat shock genes in the573

genome, the RBS Calculator was used. After running the software for 8 h, a relation 574

closer to the translation initiation rates in the genome was found, namely575

, where fxsA construction possessed the same 576

Page 24 of 44

Accep

ted

Man

uscr

ipt

24

translation initiation rate as the one used in the previous constructions (19000). These 577

newly determined RBS were then used to obtain the results gathered in Fig. 4. 578

Synthetic biology finds applications in several fields, such as biosensing (pollution 579

detection), therapeutics (drug discovery/target identification, therapeutic 580

delivery/treatment), and production of biofuels, pharmaceuticals and novel biomaterials 581

(construction/optimization of biosynthetic pathways, programming novel functionality 582

and materials). For these applications, new biological devices and systems to regulate 583

gene expression and metabolite pathways are required. Furthermore, external control 584

and triggers for such devices and systems can be designed using adequate gene 585

promoter strategies. Most promoters commonly used in heterologous protein expression 586

are inducible promoters. Chemical inducers can be expensive at laboratory or pilot-plant 587

scales and are often toxic, thus their presence in either the final product or in waste 588

effluents is highly undesirable and requires their disposal/elimination. These restrictions 589

are particularly important in food, pharmaceutical-grade proteins and other products 590

intended for human use, such as therapeutic products. Induction methods having a 591

thermo-regulated expression system with heat shock promoters are even more important 592

for in vivo applications which rely on an external stimulus to the body. The heat shock 593

promoters identified in this study can be used in the construction of new therapeutic 594

bacteria, in which the biosynthetic pathway of a therapeutic agent will be inserted and 595

whose expression will be triggered in situ by heat, emitted by laser therapy or 596

ultrasound. The expression obtained with these heat shock promoters will most likely597

not reach the same level as the one obtained with other stronger promoters such as T7; 598

however, in situ delivery will likely not require expression levels as high as those 599

necessary in bioreactors. In addition, the promoter sequences can be optimized as 600

demonstrated by the approach herein reported using RBS optimization to improve the 601

Page 25 of 44

Accep

ted

Man

uscr

ipt

25

translation efficiency and to reach the desired response and obtain the expected results 602

in the production of novel genes or pathways. Overall, a heat shock approach would be603

easier to implement and administer and will be safer since no chemicals are involved. 604

605

Acknowledgments606

This work was supported by FEDER funds through the COMPETE and ON2 program 607

and through National funds of FCT in the scope of the project FCOMP-01-0124-608

FEDER-027462, PEst-OE/EQB/LA0023/2013, NORTE-07-0124-FEDER-000028 and 609

NORTE-07-0124-FEDER-000027. Financial support for this work was provided by 610

FCT grant SFRH / BD / 51187 / 2010 and SYNBIOBACTHER project (PTDC/EBB-611

BIO/102863/2008). We acknowledge Nuno Cerca (IBB), Jörg Becker (IGC) and Rui 612

Pereira (CEB-UM) for critical discussions relative to RT-qPCR, microarrays, and heat 613

shock probes results, respectively.614

615

References616

617

Aertsen, A., Vanoirbeek, K., Spiegeleer, P., Sermon, J., Hauben, K., Farewell, A., 618

Nyström, T., Michiels, C.W., 2004. Heat shock protein-mediated resistance to high 619

hydrostatic pressure in Escherichia coli. Appl Environ Microbiol 70, 2660–2666.620

Alper, H., Fischer, C., Nevoigt, E., Stephanopoulos, G., 2005. Tuning genetic control 621

through promoter engineering. Proc. Natl. Acad. Sci. USA 102, 12678-12683.622

Anderson, J.C., Clarke, E.J., Arkin, A.P., Voigt, C.A., 2006. Environmentally controlled 623

invasion of cancer cells by engineered bacteria. J. Mol. Biol. 355, 619-627.624

Arsène, F., Tomoyasu, T., Bukau, B., 2000. The heat shock response of Escherichia 625

coli: Review. Int. J. Food Microbiol. 55, 3-9.626

Page 26 of 44

Accep

ted

Man

uscr

ipt

26

Ben-Israel, O., Ben-Israel, H., Ulitzur, S., 1998. Identification and quantification of 627

toxic chemicals by use of Escherichia coli carrying lux genes fused to stress promoters. 628

Appl Environ Microbiol. 64, 4346-4352.629

Bianchi, A.A., Baneyx, F., 1999a. Hyperosmotic shock induces the σ32 and σE stress 630

regulons of Escherichia coli. Mol. Microbiol. 34, 1029–1038.631

Bianchi, A.A., Baneyx, F., 1999b. Stress responses as a tool to detect and characterize 632

the mode of action of antibacterial agents. Appl. Environ. Microbiol. 65, 5023-5027. 633

Caspeta, L., Flores, N., Pérez, N.O., Bolívar, F., Ramírez, O.T., 2009. The effect of 634

heating rate on Escherichia coli metabolism, physiological stress, transcriptional 635

response, and production of temperature-induced recombinant protein: A scale-down 636

study. Biotechnol. Bioeng. 102, 468-482.637

Cha, H.J., Srivastava, R., Vakharia, V.N., Rao, G., Bentley, W.E., 1999. Green 638

fluorescent protein as a noninvasive stress probe in resting Escherichia coli cells. Appl.639

Environ. Microbiol. 65, 409-414.640

Chalfie, M., Tu, Y., Euskirchen, G., Ward, W.W., Prasher, D.C., 1994. Green 641

fluorescent protein as a marker for gene expression. Science 263, 802-805. 642

Chen, L.H., Emory, S.A., Bricker, A.L., Bouvet, P., Belasco, J.G., 1991. Structure and 643

Function of a Bacterial mRNA Stabilizer: Analysis of the 5' Untranslated Region of 644

ompA mRNA. J. Bacteriol. 173, 4578-4586.645

Chuang, S.E., Burland, V., Plunkett III, G., Daniels, D.L., Blattner, F.R., 1993. 646

Sequence analysis of four new heat-shock genes constituting the hslTS/ibpAB and hsl 647

VU operons in Escherichia coli. Gene 134, l-6.648

Page 27 of 44

Accep

ted

Man

uscr

ipt

27

Cormack, B.P., Valdivia, R.H., Falkow, S., 1996. FACS-optimized mutants of green 649

fluorescent protein (GFP). Gene 173, 33-38.650

Cowing, D.W., Bardwell, J.C.A., Craig, E.A., Woolford, C., Hendrix, R.W., Gross,651

C.A., 1985. Consensus sequence for Escherichia coli heat shock gene promoters. Proc.652

Natl. Acad. Sci. USA 82, 2679-2683.653

Cui, X., Churchill, G.A., 2003. Statistical tests for differential expression in cDNA 654

microarray experiments. Genome Biol. 4, 1-10.655

Dyk, T.K.V., Majarian, W.R., Konstantinov, K.B., Young, R.M., Dhurjati, P.S., 656

Larossa, R.A., 1994. Rapid and sensitive pollutant detection by induction of heat shock 657

gene-bioluminescence gene fusions. Appl Environ Microbiol. 60, 1414-1420.658

Dong, T., Kirchhof, M.G., Schellhorn, H.E., 2008. RpoS regulation of gene expression 659

during exponential growth of Escherichia coli K12. Mol. Genet. Genomics 279, 267–660

277.661

Drabner, B., Guzman, C.A., 2001. Elicitation of predictable immune responses by using 662

live bacterial vectors. Biomol. Eng. 17, 75–82.663

Dziuda, D.M., 2010. Basic analysis of gene expression microarray data, in: Dziuda, 664

D.M. (Ed.), Data Mining for Genomics and Proteomics: Analysis of Gene and Protein 665

Expression Data. John Wiley & Sons, New Jersey, pp. 17-94.666

Ellis, T., Wang, X., Collins, J.J., 2009. Diversity-based, model-guided construction of 667

synthetic gene networks with predicted functions. Nat. Biotechnol. 27, 465-471.668

Page 28 of 44

Accep

ted

Man

uscr

ipt

28

El-Samad, H., Kurata, H., Doyle, J.C., Gross, C.A., Khammash, M., 2005. Surviving 669

heat shock: Control strategies for robustness and performance. Proc. Natl. Acad. Sci.670

102, 2736-2741.671

Erickson, J.W., Vaughn, V., Walter, W.A., Neidhardt, F.C., Gross, C.A., 1987. 672

Regulation of the promoters and transcripts of rpoH, the Escherichia coli heat shock 673

regulatory gene. Gen. Dev. 1, 419-432.674

Garmory, H.S., Leary, S.E.C., Griffin, K.F., Williamson, E.D., Brown, K.A., Titball,675

R.W., 2003. The use of live attenuated bacteria as a delivery system for heterologous 676

antigens. J. Drug Target 11, 471–479.677

Genevaux, P., Georgopoulos, C., Kelley, W.L., 2007. The Hsp70 chaperone machines 678

of Escherichia coli: a paradigm for the repartition of chaperone functions. Mol.679

Microbiol. 66, 840-857.680

Georgopoulos, C., 2006. Toothpicks, serendipity and the emergence of the Escherichia 681

coli DnaK (Hsp70) and GroEL (Hsp60) chaperone machines. Genetics 174, 1699–1707.682

Gill, R.T., Cha, H.J., Jain, A., Rao, G., Bentley, W.E., 1998. Generating controlled 683

reducing environments in aerobic recombinant Escherichia coli fermentations: effects 684

on cell growth, oxygen uptake, heat shock protein expression, and in vivo CAT activity. 685

Biotechnol. Bioeng. 59, 248-259.686

Grossman, A.D., Erickson, J.W., Gross, C.A., 1984. The htpR gene product of E. coli is 687

a sigma factor for heat-shock promoters. Cell 38, 383-390.688

Guisbert, E., Herman, C., Lu, C.Z., Gross, C.A., 2004. A chaperone network controls 689

the heat shock response in E. coli. Gen. Dev. 18, 2812-2821.690

Page 29 of 44

Accep

ted

Man

uscr

ipt

29

Higgins, C.F., Causton, H.C., Dance, G.S.C., Mudd, E.A., 1993. The role of the 3’ end 691

in mRNA stability and decay, in: Belasco, J.G., Brawerman, G. (Eds.), Control of 692

messenger RNA stability. Academic Press, Inc., San Diego, pp.13-30.693

Ito, K., Akiyama, Y., 2005. Cellular functions, mechanism of action, and regulation of 694

FtsH protease. Annu. Rev. Microbiol. 59, 211-231.695

Janaszak, A., Majczac, W., Nadratowska, B., Szalewska-Palasz, A., Konopa, G., Taylor,696

A., 2007. A σ54-dependent promoter in the regulatory region of the Escherichia coli697

rpoH gene. Microbiology 153, 111-123.698

Janaszak, A., Nadratowska-Wesolowska, B., Konopa, G., Taylor, A., 2009. The P1 699

promoter of the Escherichia coli rpoH gene is utilized by σ70-RNAP or σS-RNAP 700

depending on growth phase. FEMS Microbiol. Lett. 291, 65–72.701

Jensen, P.R., Hammer, K., 1998. Artificial promoters for metabolic optimization. 702

Biotechnol. Bioeng. 58, 191-195.703

Khalil, A.S., Collins, J.J., 2010. Synthetic biology: applications come of age. Nat. Rev.704

Genet. 11, 367-379.705

Khodursky, A.B., Bernstein, J.A., Peter, B.J., Rhodius, V., Wendisch, V.F., Zimmer, 706

D.P., 2003. Escherichia coli Spotted Double-Strand DNA Microarrays: RNA 707

Extraction, Labeling, Hybridization, Quality Control, and Data Management, in: 708

Brownstein, M.J., Khodursky, A.B. (Eds.), Functional Genomics: Methods and 709

Protocols. Methods in Molecular Biology. Volume 224. Humana Press, New Jersey, pp. 710

61-78. 711

Page 30 of 44

Accep

ted

Man

uscr

ipt

30

Kobayashi, A., Hirakawa, H., Hirata, T., 2006. Growth Phase-Dependent Expression of 712

Drug Exporters in Escherichia coli and Its Contribution to Drug Tolerance. J. Bacteriol.713

188, 5693-5703.714

Kornitzer, D., Teff, D., Altuvia, S., Oppenheim, A.B., 1991. Isolation, characterization, 715

and sequence of an Escherichia coli heat shock gene, htpX. J. Bacteriol. 173, 2944-716

2953.717

Kotova, V.Y., Manukhlov, I.V., Zavilgelskii, G.B., 2010. Lux-biosensors for detection 718

of SOS-response, heat shock, and oxidative stress. Appl. Biochem. Microbiol. 46, 781-719

788.720

Kraft, M., Knüpfer, U., Wenderoth, R., Pietschmann, P., Hock, B., Horn, U., 2007. An 721

online monitoring system based on a synthetic sigma32-dependent tandem promoter for 722

visualization of insoluble proteins in the cytoplasm of Escherichia coli. Appl.723

Microbiol. Biotechnol. 75, 397-406.724

Lesley, S.A., Graziano, J., Cho, C.Y., Knuth, M.W., Klock, H.E., 2002. Gene 725

expression response to misfolded protein as a screen for soluble recombinant protein. 726

Protein Eng. 15, 153-160.727

Li, C., Wong, W.H., 2003. DNA-Chip Analyzer (dChip), in: Parmigiani, G., Garrett,728

E.S., Irizarry, R.A., Zeger, S.L. (Eds.), The analysis of gene expression data. Springer, 729

New York, pp. 120-141.730

Lindquist, S., Craig, E.A., 1988. The heat-shock proteins. Annu. Rev. Genet. 22, 631-731

677.732

Lindquist, S., 1986. The heat-shock response. Annu. Rev. Biochem. 55, 1151-1191.733

Page 31 of 44

Accep

ted

Man

uscr

ipt

31

Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using 734

real-time quantitative PCR and the 2-∆∆Ct method. Methods 25, 402-408.735

Lu, C., Albano, C.R., Bentley, W.E., Rao, G., 2005a. Quantitative and kinetic study of 736

oxidative stress regulons using green fluorescent protein. Biotechnol. Bioeng. 89, 574-737

587.738

Lu, C., Bentley, W.E., Rao, G., 2005b. Comparisons of oxidative stress response genes 739

in aerobic Escherichia coli fermentations. Biotechnol. Bioeng. 83, 864-870.740

Madigan, M.T., Martinko, J.M., Dunlap, P.V., Clark, D.P., 2009. Brock Biology of 741

Microorganisms, 12th edition. Pearson Education Ltd., San Francisco, pp. 1061.742

March, J.C., Rao, G., Bentley, W.E., 2003. Biotechnological applications of green 743

fluorescent protein. Appl. Microbiol. Biotechnol. 62, 303-315.744

Messaoudi, N., Gautier, V., Kthiri, F., Lelandais, G., Mihoub, M., Joseleau-Petit, D., 745

Caldas, T., Bohn, C., Tolosa, L., Rao, G., Tao, K., Landoulsi, A., Bouloc, P., Richarme,746

G., 2013. Global stress response in a prokaryotic model of DJ-1-associated 747

Parkinsonism. J. Bacteriol. 195, 1167-1178.748

Nagai, H., Yuzawa, H., Yura, T., 1991. Interplay of two cis-acting mRNA regions in 749

translational control of σ32 synthesis during the heat shock response of Escherichia coli. 750

Proc. Natl. Acad. Sci. USA 88, 10515-10519.751

Nemecek, S., Marisch, K., Juric, R., Bayer, K., 2008. Design of transcriptional fusions 752

of stress sensitive promoters and GFP to monitor the overburden of Escherichia coli753

hosts during recombinant protein production. Bioprocess Biosyst. Eng. 31, 47-53.754

Page 32 of 44

Accep

ted

Man

uscr

ipt

32

Nonaka, G., Blankschien, M., Herman, C., Gross, C.A., Rhodius, V.A., 2006. Regulon 755

and promoter analysis of the E. coli heat-shock factor, σ32, reveals a multifaceted 756

cellular response to heat stress. Gen. Dev. 20, 1776-1789.757

Pérez, J.M., Calderón, I.L., Arenas, F.A., Fuentes, D.E., Pradenas, G.A., Fuentes, E.L., 758

Sandoval, J.M., Castro, M.E., Elías, A.O., Vásquez, C.C., 2007. Bacterial Toxicity of 759

Potassium Tellurite: Unveiling an Ancient Enigma. PLoS ONE e211, 1-9.760

Raina, S., Missiakas, D., Georgopoulos, C., 1995. The rpoE gene encoding the σE (σ24) 761

heat shock sigma factor of Escherichia coli. EMBO J. 14, 1043-1055.762

Rasouly, A., Schonbrun, M., Shenhar, Y., Ron, E.Z., 2009. YbeY, a heat Shock protein 763

involved in translation in Escherichia coli. J. Bacteriol. 191, 2649–2655.764

Richmond, C.S., Glasner, J.D., Mau, R., Jin, H., Blattner, F.R., 1999. Genome-wide 765

expression profiling in Escherichia coli K-12. Nucleic Acids Res. 27, 3821-3835.766

Rodrigues, L.R., Kluskens, L.D., 2011. Synthetic Biology & Bioinformatics: Prospects 767

in the cancer arena. In: Lopes, H.S., Cruz, L.M., editors. Computational Biology and 768

Applied Bioinformatics. InTech, Rijeka, pp. 159-186. 769

Ross, W., Gosink, K.K., Salamon, J., Igarashi, K., Zou, C., Ishihama, A., Severinov, K., 770

Gourse, R.L., 1993. A Third Recognition Element in Bacteria Promoters: DNA Binding 771

by the α Subunit of RNA Polymerase. Science 262, 1407-1413.772

Rozen, S., Skaletsky, H.J., 2000. Primer3 on the WWW for general users and for 773

biologist programmers. In: Krawetz, S., Misener, S., editors. Bioinformatics Methods 774

and Protocols: Methods in Molecular Biology. Humana Press, New Jersey, pp. 365-386.775

Page 33 of 44

Accep

ted

Man

uscr

ipt

33

Sagi, E., Hever, N., Rosen, R., Bartolome, A.J., Premkumar, J.R., Ulber, R., Lev, O., 776

Scheper, T., Belkin, S., 2003. Fluorescence and bioluminescence reporter functions in 777

genetically modified bacterial sensor strains. Sens. Actuators B 90, 2-8.778

Salis, H.M., 2011. The Ribosome Binding Site Calculator. Meth. Enzymol. 498, 19-42.779

Seo, J.H., Kang, D.G., Cha, H.J., 2003. Comparison of cellular stress levels and green-780

fluorescent-protein expression in several Escherichia coli strains. Biotechnol. Appl.781

Biochem. 37, 103-107.782

Shankar, S., Pillai, M.R., 2011. Translating cancer research by synthetic biology. Mol. 783

BioSyst. 7, 1802-1810.784

Shin, D., Lim, S., Seok, Y.J., Ryu, S., 2001. Heat shock RNA polymerase (Eσ32) is 785

involved in the transcription of mlc and crucial for induction of the Mlc regulon by 786

glucose in Escherichia coli. J. Biol. Chem. 276, 25871–25875.787

Tilly, K., Spence, J., Georgopoulos, C., 1989. Modulation of stability of the Escherichia 788

coli heat shock regulatory factor σ32. J. Bacteriol. 171, 1585-1589.789

Tsui, H.C.T., Feng, G., Winkler, M.E., 1996. Transcription of the mutL repair, miaA 790

tRNA modification, hfq pleiotropic regulator, and hflA region protease genes of 791

Escherichia coli K-12 from clustered Eσ32-specific promoters during heat shock. J.792

Bacteriol. 178, 5719–5731.793

Valdez-Cruz, N.A., Caspeta, L., Pérez, N.O., Ramírez, O.T., Trujillo-Roldán, M.A.,794

2010. Production of recombinant proteins in E. coli by the heat inducible expression 795

system based on the phage lambda pL and/or pR promoters. Microb. Cell Fact. 9, 1-16.796

Page 34 of 44

Accep

ted

Man

uscr

ipt

34

Van Dyk, T.K., Majarian, W.R., Konstantinov, K.B., Young, R.M., Dhurjati, P.S., 797

LaRossa, R.A., 1994. Rapid and sensitive pollutant detection by induction of heat shock 798

gene-bioluminescence gene fusions. Appl. Environ. Microbiol. 60, 1414-1420.799

Wade, J.T., Roa, D.C., Grainger, D.C., Hurd, D., Busby, S.J.W., Struhl, K., Nudler, E.,800

2006. Extensive functional overlap between σ factors in Escherichia coli. Nat. Struct.801

Mol. Biol. 13, 806-814.802

Yamamori, T., Ito, K., Nakamura, Y., Yura, T., 1978. Transient regulation of protein 803

synthesis in Escherichia coli upon shift-up of growth temperature. J. Bacteriol. 134, 804

1133-1140.805

Yura, T., Nakanigashi, K., 1999. Regulation of the heat-shock response. Curr. Opin.806

Microbiol. 2, 153-158.807

Zhao, K., Liu, M., Burgess, R.R., 2005. The global transcriptional response of 808

Escherichia coli to induced σ32 protein involves σ32 regulon activation followed by 809

inactivation and degradation of σ32 in vivo. J. Biol. Chem. 280, 17758-17768.810

811

812

813

Page 35 of 44

Accep

ted

Man

uscr

ipt

35

813

814

Figure Captions815

Fig. 1. Roadmap of the microarrays data treatment (A) and the Spearman correlation 816

matrix and hierarchical clustering of 189 genes whose expression is significantly altered 817

following heat shock induction (B). The red colour in the hierarchical clustering 818

represents expression levels above the average of gene expression in all samples, the 819

white colour represents an average expression, and the blue colour represents expression 820

levels below average.821

Fig. 2. Gene expression profiles (A) and fold change (B) during the temperature up-shift 822

from 37 to 45ºC in flask.823

Fig. 3. Gene expression profiles (A, B) and fold changes (C, D) during up-shift from 37 824

to 45 ºC (A, C) and 30 to 50ºC (B, D) and subsequent down-shift in bioreactor. 825

Fig. 4. Specific fluorescence intensity after heat shock in E. coli MG1655 DE3 826

transformed with pETduet_GFPmut3b plasmids with different promoters: ibpA827

promoter (), dnaK promoter () and fxsA promoter () (control – 37 or 30ºC), ibpA828

promoter (■), dnaK promoter (●) and fxsA promoter (▲) (heat shock – 45 or 50ºC). 829

The three biological replicates for each experiment were performed in different days.830

The fluorescence values in the figures on the right correspond to the maximum values 831

obtained for each case (1 h after heat shock induction). 832

833834

Page 36 of 44

Accep

ted

Man

uscr

ipt

36

Table 1. Set of primers for PCR amplification used in GFP and heat shock promoters 834

for stress probe construction (forward and reverse primers – fw and rev)835

Primer Name Primer sequence (5’ > 3’)GFPmut3b.fw AAAAAGGATCCAATGCGTAAAGGAGAAGAACTGFPmut3b.rev AAAAAGGCGCGCCTTATTATTTGTATAGTTCATCCATGCCibpA_P_fw AAAAAGCATGCGAGCTCAAAATAACATCATCATTACGTCGibpA_P_rev AAAAAGGATCCTCCTTAATAGTTCGCTCGTACGATGTAAAAATGGGTCTGdnaK_P_fw AAAAAGCATGCGAGCTCATGCCTTGGCTGCGATTdnaK_P_rev AAAAAGGATCCTCCTCGTCTATTCCGACTCCTATTCGGTCATCATGTGGTfxsA_P_fw AAAAAGCATGCGAGCTCGAGGATTTCTACCGTAATCTGfxsA_P_rev AAAAAGGATCCTCCCGATCCGTCGTGCTTTCTTTAGCTTATATTGTGGTCATTAG

836

837

Page 37 of 44

Accep

ted

Man

uscr

ipt

37

Table 2. Set of primers used in RT (reverse primers - rev) and qPCR amplification 837

(forward and reverse primers – fw and rev)838

Gene name Primer name Primer Sequence (5’ > 3’)

dnaK.fw AATCGAACTGTCTTCCGCTC-dnaK

dnaK.rev TTCGCACGAGTCACTTTGATibpA.fw ACAAAAAGAGCGCACCTATCT

ibpAibpA.rev CAGGTTAGCACCACGAACAfxsA.fw CGGCGGAGATGATTAAAAGT

fxsAfxsA.rev CGGCGGCAATAAAAGTAGA16S.fw TAGCGGTGAAATGCGTAGAG

rrsA - 16S16S.rev TAATCCTGTTTGCTCCCCAC

839

840

841

Page 38 of 44

Accep

ted

Man

uscr

ipt

38

Table 3. Heat shock genes induction ratio, expression values (average of two biological 841replicates) and t and S test results. Genes marked with "*" represent the ones that are 842directly downstream of the σ32 factor dependent promoters. The list presented is ordered843according to the average rate of induction (Test/Control) of the genes. q-value was 0 for 844all the genes. In bold it is possible to observe the promoters chosen for posterior 845validation.846

Blattner Induction Ratios Expression values t-test - dChipS-test -SAMGene code

Number

Gene

Test/Control Control Test t p- value S

1762978_s_at b3686 ibpB 224.76 22.55 5071.54 8.25 0.08 8.39

1769060_s_at b3687 ibpA* 55.68 106.34 5917.71 8.95 0.07 9.05

1765041_s_at b1322 ycjF 23.82 62.22 1473.74 14.38 0.04 13.55

1768270_s_at b1321 ycjX* 21.77 102.70 2230.95 24.86 0.03 24.53

1766244_s_at b3635 mutM* 21.1 42.90 907.10 5.10 0.12 4.92

1762193_s_at b3685 yidE 21.09 21.95 463.43 2.29 0.26 2.23

1765416_s_at b1380 ldhA* 20.12 113.35 2282.35 15.42 0.04 14.83

1760394_s_at b1594 mlc* 18.78 70.46 1324.72 3.83 0.16 3.76

1763446_s_at b4140 fxsA* 16.89 104.78 1742.74 95.73 0.01 134.67

1761819_s_at b0473 htpG* 13.02 347.16 4469.62 27.21 0.02 28.35

1766513_s_at b0492 ybbN* 11.15 215.08 2397.06 7.33 0.09 7.18

1761806_s_at b3401 hsp33 9.89 138.25 1365.18 8.22 0.08 7.93

1760965_s_at b1829 htpX* 9.82 379.64 3742.23 16.26 0.04 15.99

1763516_s_at b1060 yceP* 9.44 224.83 2117.17 6.75 0.09 6.60

1764871_s_at b2592 clpB* 9.17 548.94 4955.68 48.22 0.01 679.74

1761621_s_at b3498 prlC* 9.17 176.66 1609.55 10.99 0.06 10.54

1763737_s_at b3594 yibA* 9.01 41.95 376.25 3.78 0.16 3.54

1761422_s_at b3400 yrfH* 8.66 292.95 2517.09 10.39 0.06 10.14

1768615_s_at b3497 yhiQ 8.29 135.89 1127.29 5.65 0.11 5.46

1764204_s_at b3292 zntR 8.24 187.82 1541.54 72.42 0.01 155.46

1766949_s_at b3932 hslV* 8.07 416.46 3351.77 44.71 0.01 46.65

1768518_s_at b4143 groEL 7.90 743.13 5831.04 37.87 0.02 55.53

1763357_s_at b0014 dnaK* 7.59 814.09 6120.95 24.91 0.03 28.07

1769019_s_at b0015 dnaJ 7.08 559.61 3946.50 23.26 0.03 23.91

1767479_s_at b1593 ynfK 7.08 35.13 246.56 4.01 0.16 3.60

1764894_s_at b4142 groES* 6.79 772.53 5188.85 25.73 0.03 148.96

1767623_s_at b0439 lon* 6.76 604.83 4079.96 18.37 0.04 19.04

1762240_s_at b3931 hslU 6.58 518.63 3370.51 24.98 0.03 30.41

1764710_s_at b3293 yhdN* 6.42 369.45 2377.17 38.28 0.02 40.39

1767508_s_at b1814 sdaA* 6.37 536.85 3429.90 7.25 0.09 7.13

1764062_s_at b1379 hslJ 6.30 162.79 1016.68 21.90 0.03 19.99

1768300_s_at b0016 yi81_1 5.74 374.36 2133.52 23.63 0.03 22.85

1768989_s_at b1280 yciM 5.64 217.00 1218.50 18.14 0.04 17.24

1759242_s_at b0630 lipB 5.61 220.41 1235.10 64.04 0.01 70.13

Page 39 of 44

Accep

ted

Man

uscr

ipt

39

847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887

Highlights888

889

Heat shock promoters from Escherichia coli were identified and validated by microarrays and 890

RT-qPCR, respectively891

Page 40 of 44

Accep

ted

Man

uscr

ipt

40

ibpA, dnaK and fxsA were found to be the strongest heat shock promoters possessing distinct 892

promoter strengths893

Stress probes designed with those promoters suggest their potential for controlling gene 894

expression895

It is crucial to test and optimize each promoter to obtain translational efficiency that balances 896

transcription levels 897

Escherichia coli heat shock promoters can be explored as synthetic biology tools to trigger gene 898

expression 899

900

901

902

903

Page 41 of 44

Accep

ted

Man

uscr

ipt

TA ~ TB

expression

Underexpressed genes

Overexpressed genes

TB expression

much higher to

TA expression

Higher

expression

in TA

TB expression

higher or equal

(A)

128 genes 56 genes

189 genes

703 genes

4070 genes

Selection of

E. coli K-12

MG1655 strain

genes

Filtering Criteria

Comparison

Criteria

Selection of

significant genes

by t- test

Selection of

significant genes

by S-test

(B)

Figure 1

Page 42 of 44

Accep

ted

Man

uscr

ipt

1

10

100

1000

10000

Heat Shock Gene

Fo

ld C

ha

ng

e

dnaK

ibpA

fxsA

1

10

100

1000

10000

100000

Control - 37 ºC Heat Shock - 45 ºC

Ex

pre

ssio

n L

evel

1

07

dnaK

ibpA

fxsA

(A) (B)

Figure 2

Page 43 of 44

Accep

ted

Man

uscr

ipt

(C) (D)

(B) (A)

1

10

100

1000

10000

100000

M N O P Q R S T U V W

Ex

pre

ssio

n l

evel

x 1

07

dnaK

ibpA

fxsA

1

10

100

1000

10000

M N O P Q R S T U V W

Fo

ld C

ha

ng

e

dnaK

ibpA

fxsA

0

1

10

100

1000

10000

A B C D E F G H I J K L

Fo

ld C

ha

ng

e

dnaK

ibpA

fxsA

1

10

100

1000

10000

100000

A B C D E F G H I J K L

Ex

pre

ssio

n l

evel

x 1

07

dnaKibpAfxsA

Figure 3

Page 44 of 44

Accep

ted

Man

uscr

ipt

Sp

eci

fic

Flu

ores

cen

ce I

nte

nsi

ty (

Flu

ore

scen

ce I

nte

nsi

ty O

D-1

) *

10

-3

0

5

10

15

20

25

37 ºC 45 ºC 37 ºC 50 ºC 30 ºC 50 ºC

Heat Shock Experiments

dnaK promoter

Control

Heat Shock

0

5

10

15

20

25

0 2 4 6

Time after induction (h)

37ºC - 50 ºC, 10 min

0

1

2

3

4

5

6

37 ºC 45 ºC 37 ºC 50 ºC 30 ºC 50 ºC

Heat Shock Experiments

ibpA promoter

ControlHeat Shock

0

2

4

6

8

10

12

14

0 2 4 6Time after induction (h)

37 ºC - 45 ºC, 5 min

0

2

4

6

8

10

12

14

0 2 4 6

Time after induction (h)

30 ºC - 50ºC, 10 min

0,0

0,1

0,2

0,3

0,4

0,5

0,6

37 ºC 45 ºC 37 ºC 50 ºC 30 ºC 50 ºC

Heat shock Experiments

fxsA promoter

Control

Heat Shock

Figure 4